I do not see that as a problem; I've had reasonable luck running SJava
visualization code from Orca and ROrca remotely (over 1000bT), and VNC
works as well for that, but you are right, have to pay attention to
screen color depth.
Now whether SJava can be used for a stable system is a completely
separate matter; I have had awful luck with it (i.e. semi-stable
rather than rock-solid), which is why we've moved to C++/OpenGL
instead of extending Orca (and hence Java/Java3d) for our
visualization/virtual reality systems.
Hopefully Simon has or will solve that problem; I have confidence
:-).
best,
-tony
"Liaw, Andy" <andy_liaw@merck.com> writes:
I just have one question. From my very little (and negative) experience
with the Java GUI for S-PLUS on *nix, one big problem is that it doesn't
work when running S-PLUS remotely, either via X or VNC (unless the display
is 8-bit): The GUI is running, but on screen I just see one big black
rectangle. Would this be a problem for R as well? My impression was that
that is totally the problem with Java. (BTW, I believe S-PLUS also installs
and uses its private JRE.)
Best,
Andy
From: Simon Urbanek
On Feb 25, 2004, at 4:52 PM, Prof Brian Ripley wrote:
That's very interesting, but as you know Java installation
problematic. (It looks like the current Sun JVM 1.4.2_03
issues, including not working with the R search engine, and
on
mozilla.org about its Windows plugin.)
I'm aware of the problems caused by various JVM versions.
Although our
Java code is not directly affected, we plan to supply an optional
installer (for Windows) which will take care of the JVM and
install an
appropriate private JRE if necessary.
We bundled a version of Tcl/Tk with the R Windows binary to try to
avoid
compatibility issues, and even then I would prefer a native Windows
solution for this one.
I'm definitely not against a native editor. I don't want to
discourage
Chris, especially if he has already a working prototype. To the
contrary - one implicit point of my post was that we should stay in
contact, since we have implemented an editor as well and
thought about
the interface and the features, so we could exchange our experiences.
I also wanted to say that we may want to talk about bigger
scope here.
My concern is that the number of scattered GUIs for various platforms
is increasing and each GUI has own interfaces, inconsistent with
others, making it hard to switch between GUIs. If one wants to add
functionality to the 'preferred' GUI then others don't have
any benefit
from the extension. A platform-independent GUI would solve this
problem, since an improvement would affect all platforms. Furthermore
using Java, it is possible to offer the quite powerful Java
widget set
to R users, seamlessly integrated into the GUI, thus allowing
users to
extend the GUI as they please - using R. (Besides, users could even
integrate their own Java code with R, but that's another
story...). By
now we also have an own, working Java graphics device allowing us to
achieve really tight integration of all aspects of R. I don't claim
that this should replace the Windows GUI, but I think it might be an
interesting, flexible alternative. I know that the project is all but
easy, nevertheless I believe that we got quite far by now.
Simon
---
Simon Urbanek
Department of computer oriented statistics and data analysis
University of Augsburg
Universit?tsstr. 14
86135 Augsburg
Germany
Tel: +49-821-598-2236
Fax: +49-821-598-2200
Simon.Urbanek@Math.Uni-Augsburg.de
http://simon.urbanek.info